An overview of the biased assignment design is presented. If properly used, the biased assignment design will preserve internal validity and, unlike random design, will allow certain of the criteria deemed important by judges and correctional officials to guide the assignment of offenders to groups. The major feature of this design is the assignment to conditions on the basis of a cutoff on some quantified measurement. The four steps of using the biased assignment design are presented. Step one involves obtaining data on the criteria underlying the formation of comparison groups. Step two is where the distribution of scores on the pretest is inspected and an exact cutoff point for assignment to either the control or experimental group is established. Step three is the crucial step for preserving the internal validity of the evaluation. Each offender should be assigned a group membership solely on the basis of his or her ranking per the pretest cutoff value. Step four is program implementation. The article concludes with a consideration of the analysis of the pretest and posttest data. 5 notes and 21 references (Author abstract modified)